Twin and adoption studies have found evidence for genetic influences in substance abuse, but studies differ considerably in the amount of variation attributed to genetic and environmental factors. Since the studies differ in the instruments they used to assess substance abuse, as well as in other aspects, it is impossible to use the results of these studies to establish which measuring instruments are most sensitive in estimating genetic contributions. Using data from the Minnesota Twin Registry, we are comparing intrapair correlations in substance use by monozygotic and dizygotic twins on commonly used assessment instruments to determine which instruments yield the highest heritability estimates. Measures include DSM III criteria, Feighner and Cloninger classifications, and frequency, quantity and density of use, including heavy use and problemmatic use. Instruments that are the most sensitive in estimating genetic influences will have the best chance to distinguish genetically based forms of alcoholism from those that are predominantly environmentally based. Being able to select individuals and families with large genetic influences will enhance the changes of success for molecular genetic approaches to find the genes that are important in substance abuse.